Re: DeadLocks - Mailing list pgsql-general

From Gordon Campbell
Subject Re: DeadLocks
Date
Msg-id 07c201c1268c$85a5ff50$cc213b80@gordie
Whole thread Raw
In response to DeadLocks  ("Gordon Campbell" <gordie@columbia.edu>)
List pgsql-general
1. I'll upgrade asap.
2. Yeah, I pretty sure it's a deadlock or as Tom says a connection that is
not getting cleaned up
3. This is the hard part, I can't replicate. I could run update x times in a
row using the same key and it will work, but on eventually I'll get shut
out. At the moment, I'm the only one hitting this table/db.

Thanks for you feedback.


----- Original Message -----
From: "Joseph Shraibman" <jks@selectacast.net>
To: "Gordon Campbell" <gordie@columbia.edu>
Sent: Tuesday, August 14, 2001 5:35 PM
Subject: Re: [GENERAL] DeadLocks


> I have had some deadlocks, but haven't been able to reproduce them
reliably.
>
> 1) Use the lastest version. 7.0.3 is old and the maintainers won't want
> to debug it
> 2) Are you *sure* it is deadlocking?  is the system just really slow?
> 3) Can you make a test case that always locks?
>
> Gordon Campbell wrote:
> > This is my first posting to this site, but it's my last resort. We're
> > running 7.0.3. I have 20 servlets that make use of our postgres
> > database. All of the programs are able to perform their assigned SQL
> > tasks, but not on a consistent basis. Often and unpredictably, different
> > programs performing different functions (add, delete, update) will
> > essentially hang. I will view the process (ps -ef) and postgres will be
> > in an (commit, delete or update) state but does not finish. All others
> > wishing to perform a task either via the browser or at command line
> > (psql) will hang. The only way I can free up the process is having our
> > sysadmin "kill" the process.
> >
> >
> >
> > The ten tables contain less than 400 rows. We are in test mode and will
> > go into production after this "show stopper" is resolved.
> >
> >
> >
> > I have run "vacuum" which also hangs on one table that contain 382 rows.
> > I am the sole tester, so I am not clashing with other persons. The
> > servlets each make their own connection using the postgres' jdbc driver.
> > The connections are closed at the end of the program. These programs are
> > very simple insert, update, select statements.
> >
> >
> >
> > Has anybody else had this problem? What debugging methods/tools/logs
> > will help.
> >
> >
> >
> > This is a very small development shop, so I'm throwing this out to a
> > larger community for the first time.
> >
> >
> >
> > Thanks,
> >
> > Gordon Campbell
> >
> > gordie@columbia.edu <mailto:gordie@columbia.edu>
> >
> > Educational Technologist
> >
> > 212-854-1869
> >
>
>
> --
> Joseph Shraibman
> jks@selectacast.net
> Increase signal to noise ratio.  http://www.targabot.com
>


pgsql-general by date:

Previous
From: newsreader@mediaone.net
Date:
Subject: Re: unix timestamp
Next
From: Tom Lane
Date:
Subject: Re: Re[2]: Perfomance decreasing